Date of Award
4-2011
Degree Name
Master of Arts
Department
Geography
First Advisor
Dr. Kathleen M. Baker
Second Advisor
Dr. Charles Emerson
Third Advisor
Dr. Lisa DeChano-Cook
Access Setting
Masters Thesis-Campus Only
Abstract
The last 50 years has seen a remarkable growth in potato production worldwide. Coupled with this growth is the resurgence in Phytophthora infestans (Mont.) De Bary, the cause of potato late blight. Weather-based disease forecasting systems are vital to the mitigation of outbreaks of potato late blight. Artificial neural networks have proved adept at accurately performing the task of utilizing the large volumes of real-time weather data archived as Unedited Local Climatology Data. The inherent data-driven nature of artificial neural networks limits the spatial reach of forecasts beyond the source of the data. This research focuses on the analysis of basic non-parametric interpolation models for artificial neural network solutions at 45 Unedited Local Climatology Data stations in the lower peninsula of Michigan. The effectiveness and efficiency of using artificial neural network solutions in weather-based disease forecasting systems will be discussed, using potato late blight as a case study.
Recommended Citation
Rivet, Douglas, "Non-Parametric Interpolation and Weather Research Forecast Models for Potato Late Blight Risk Forecasting" (2011). Masters Theses. 4250.
https://scholarworks.wmich.edu/masters_theses/4250